> -----Original Message-----
> From: Horia Geanta
> Sent: Monday, July 22, 2019 7:55 PM
> To: Vakul Garg <vakul.garg@xxxxxxx>; linux-crypto@xxxxxxxxxxxxxxx
> Cc: Aymen Sghaier <aymen.sghaier@xxxxxxx>;
> herbert@xxxxxxxxxxxxxxxxxxx
> Subject: Re: [PATCH v2] crypto: caam/qi2 - Add printing dpseci fq stats using
> debugfs
>
> On 7/19/2019 2:23 PM, Vakul Garg wrote:
> [...]
> > +if CRYPTO_DEV_FSL_DPAA2_CAAM
> > +
> > +config CRYPTO_DEV_FSL_DPAA2_CAAM_DEBUGFS
> > + depends on DEBUG_FS
> > + bool "Enable debugfs support"
> > + help
> > + Selecting this will enable printing of various debug information
> > + in the DPAA2 CAAM driver.
> > +
> > +endif
> Let's enable this based on CONFIG_DEBUG_FS.
>
> > diff --git a/drivers/crypto/caam/Makefile
> > b/drivers/crypto/caam/Makefile index 9ab4e81ea21e..e4e9fa481a44
> 100644
> > --- a/drivers/crypto/caam/Makefile
> > +++ b/drivers/crypto/caam/Makefile
> > @@ -30,3 +30,4 @@ endif
> > obj-$(CONFIG_CRYPTO_DEV_FSL_DPAA2_CAAM) += dpaa2_caam.o
> >
> > dpaa2_caam-y := caamalg_qi2.o dpseci.o
> > +dpaa2_caam-$(CONFIG_CRYPTO_DEV_FSL_DPAA2_CAAM_DEBUGFS) +=
> > +dpseci-debugfs.o
> dpaa2_caam-$(CONFIG_DEBUG_FS)
>
> [...]
> > diff --git a/drivers/crypto/caam/caamalg_qi2.h
> > b/drivers/crypto/caam/caamalg_qi2.h
> > index 973f6296bc6f..b450e2a25c1f 100644
> > --- a/drivers/crypto/caam/caamalg_qi2.h
> > +++ b/drivers/crypto/caam/caamalg_qi2.h
> > @@ -10,6 +10,7 @@
> > #include <soc/fsl/dpaa2-io.h>
> > #include <soc/fsl/dpaa2-fd.h>
> > #include <linux/threads.h>
> > +#include <linux/netdevice.h>
> How is this change related to current patch?
>
> > #include "dpseci.h"
> > #include "desc_constr.h"
> >
> > @@ -64,6 +65,7 @@ struct dpaa2_caam_priv {
> > struct iommu_domain *domain;
> >
> > struct dpaa2_caam_priv_per_cpu __percpu *ppriv;
> > + struct dentry *dfs_root;
> dfs_root is used only in dpseci-debugfs.c, let's have it there as global.
>
I submitted this change in v3. There is still a minor issue with this patch version.
Before submitting the next v4, I have a question.
Could there be a situation that there are multiple dpseci objects assigned to kernel?
In that case, we need to maintain dfs_root for each separately.
> Horia
